18F-FDG-PET hypometabolism as a predictor of favourable outcome in epilepsy surgery: protocol for a systematic review and meta-analysis.

INTRODUCTION: A substantial proportion of patients who undergo surgery for drug resistant focal epilepsy do not become seizure free. While some factors, such as the detection of hippocampal sclerosis or a resectable lesion on MRI and electroencephalogram-MRI concordance, can predict favourable outcomes in epilepsy surgery, the prognostic value of the detection of focal hypometabolism with 18F-fluorodeoxyglucose positive emission tomography (18F-FDG-PET) hypometabolism is uncertain. We propose a protocol for a systematic review and meta-analysis to examine whether localisation with 18F-FDG-PET hypometabolism predicts favourable outcomes in epilepsy surgery. METHODS AND ANALYSIS: A systematic literature search of Medline, Embase and Web of Science will be undertaken. Publications which include evaluation with 18F-FDG-PET prior to surgery for drug resistant focal epilepsy, and which report ≥12 months of postoperative surgical outcome data will be included. Non-human, non-English language publications, publications with fewer than 10 participants and unpublished data will be excluded. Screening and full-text review of publications for inclusion will be undertaken by two independent investigators, with discrepancies resolved by consensus or a third investigator. Data will be extracted and pooled using random effects meta-analysis, with heterogeneity quantified using the I2 analysis. ETHICS AND DISSEMINATION: Ethics approval is not required. Once complete, the systematic review will be published in a peer-reviewed journal. PROSPERO REGISTRATION NUMBER: CRD42022324823. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This protocol has been devised in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses protocols guidelines.

Publication of the protocol provides transparency regarding the methodology of the systematic review, reduces the risk of meta-bias and avoids duplication.

We anticipate that most publications addressing the review question will be non-randomised, which are inherently associated with a high risk of bias. We intend to evaluate each included publication for bias, using the Newcastle-Ottawa Scale for non-randomised studies and the Cochrane Risk of Bias 2 tool for randomised studies.

Relevant data may be missed by excluding non-English language publications

Introduction

Surgical resection is an efficacious and safe treatment for selected patients with drug resistant focal epilepsy,1–3 and is also associated with higher rates of seizure freedom compared with continued best medical therapy.4 Seizure freedom is a widely used epilepsy surgery outcome measure and is a strong predictor of improvement in health-related quality of life.5 However, despite rigorous patient selection practices, approximately one third of patients who undergo epilepsy surgery do not become seizure free.6–8 Identifying predictive factors of favourable outcome in epilepsy surgery may improve the proportion of patients achieving seizure freedom by better informing the patient selection process.

Successful epilepsy surgery relies on the accurate identification and resection of the epileptogenic zone. Brain 18F-fluorodeoxyglucose positive emission tomography (18F-FDG-PET) measures regional cerebral glucose uptake and is a marker of neuronal cellular activity. Hypometabolism is an important abnormal finding in interictal 18F-FDG-PET and reflects dysfunctional brain tissue. It is commonly used in combination with other invasive and non-invasive methods of identifying the epileptogenic zone, such as MRI, ictal scalp electroencephalogram (EEG) and stereotactic EEG, to formulate hypotheses regarding epileptogenic zone localisation, which in turn informs surgical planning. Several factors, for example, the detection of hippocampal sclerosis8 9 or a resectable lesion on MRI,7–10 and concordant MRI and ictal EEG abnormalities,8 9 11 12 have been shown to be consistently predictive of favourable outcomes for patients following epilepsy surgery.13 Previous meta-analyses addressing the role of localisation with 18F-FDG-PET as a predictor of epilepsy surgery outcome have focused on temporal14 15 or frontal10 lobe epilepsy surgery with inconsistent results. However, these meta-analyses were underpowered, and only one provided a detailed analysis of the role of 18F-FDG-PET. Furthermore, currently 18F-FDG-PET hypometabolism extending beyond a single brain lobe may be considered ‘not useful’ in the decision-making process regarding epilepsy surgery,16 however, the presence of focal (involving a single lobe), compared with regional (involving two adjacent lobes) or diffuse (extending beyond two adjacent lobes) hypometabolism may be of prognostic importance. We propose a protocol for a systematic review and meta-analysis designed to examine the primary question: does localisation with 18F-FDG-PET hypometabolism predict favourable outcomes in surgery for drug resistant focal epilepsy? Secondary questions that will also be addressed include:

Is the extent of 18F-FDG-PET hypometabolism, for example focal, regional or diffuse, associated with differences in outcome following surgery for drug resistant focal epilepsy?

Does a certain proportion of the 18F-FDG-PET hypometabolism region need to be resected to achieve seizure freedom, and does this differ between brain lobes?

Do differences in the underlying pathology on histological examination of the resected tissue impact localisation with 18F-FDG-PET hypometabolism, and is this associated with differences in outcome following surgery for drug resistant focal epilepsy?

Methods and analysis

Protocol and registration

This protocol was written in accordance with thePreferred Reporting Items for Systematic Reviews and Meta-Analyses protocols guidelines17 (online supplemental table 1). The protocol is registered in PROSPERO (CRD42022324823).

Population

The population are patients of all ages who have undergone resective surgery (lesionectomy or lobectomy) for drug resistant focal epilepsy.

Intervention assessed

The intervention to be assessed is localising 18F-FDG-PET hypometabolism, as defined by concordance with other diagnostic methods (MRI, ictal scalp EEG, stereotactic EEG and/ or final decision regarding surgical site alone) and ipsilateral to the surgical site.

Control population

The control population will be individuals who undergo resective surgery for drug resistant focal epilepsy who have non-localising or no 18F-FDG-PET hypometabolism.

Outcome

We will include Engel or International League Against Epilepsy (ILAE) outcome classification systems, which are widely used measures of epilepsy surgery outcomes. Favourable outcome will be defined as Engel class I ‘free of disabling seizures,’ ILAE class 1 ‘completely seizure free; no auras’ or ILAE class 2 ‘only auras; no other seizures’.18

Eligibility criteria

The systematic review will include publications with ≥10 participants who were preoperatively evaluated with 18F-FDG-PET prior to resective surgery for drug resistant focal epilepsy. In addition, only studies with ≥12 months of postsurgical follow-up and reporting seizure outcome data will be included.

Review articles, letters, unpublished data, non-human studies and publications in languages other than English will be excluded from the systematic review. Patients who have undergone non-resective surgeries, or procedures primarily performed with palliative intent, including hemispherectomy, will be excluded. Studies in which the surgical outcome is unable to be correlated with the 18F-FDG-PET finding(s) will also be excluded.

Search strategy

Three electronic databases will be searched for eligible publications: Medline (Ovid), Embase (Ovid) and Web of Science (all databases), with the initial database searches occurring on 3 May 2022. The search terms include expanded forms and variations on “epilepsy”, “seizure”, “neurosurgery”, “positron emission tomography” and “functional neuroimaging”. The initial search will not be filtered for English language or publication type. The full search strategy for each database is available in online supplemental file.

The reference list of all included studies will be screened for other eligible publications that have not already been screened.

Selection process

Retrieved studies from database searches will be managed using Covidence Software (Veritas Health Innovation, Melbourne, Australia). All retrieved studies will be screened using title and abstract by two independent investigators. Any publication considered to be potentially eligible for inclusion by one or both investigators will be included for full-text review. The full-text reviews will be undertaken independently by two investigators, and any disagreements regarding eligibility for inclusion will be resolved by a third investigator.

Data collection

The data to be extracted includes publication details (publication year, first author, author affiliations, title, journal), patient demographics, epilepsy characteristics (seizure type, seizure frequency, age of epilepsy onset), 18F-FDG-PET characteristics, method of 18F-FDG-PET interpretation, presurgical investigations other than 18F-FDG-PET, histopathology, proportion of hypometabolism zone resected, surgical outcome and length of postoperative follow-up.

Data will be extracted by two independent investigators into a predefined data extraction spreadsheet. We intend to contact the authors for further information if the relevant data is not reported in the original publication. The data extraction results will be compared after the first 10 publications, and if congruent, the remainder of the data collection will be performed by a single investigator. Discrepancies will be resolved by discussion and/ or by a third investigator.

Risk of bias assessment

All included studies in our systematic review and meta-analysis will be assessed for bias independently by two reviewers, and discrepancies will be resolved by consensus. Non-randomised studies, including case–control and cohort studies, will be assessed for bias using the Newcastle-Ottawa Scale (NOS).19 The NOS assesses eight items across three domains: selection, comparability and outcome. Each item is assessed, and stars are awarded for high quality according to the NOS guidelines. The maximum score is four stars for selection, two stars for comparability and three stars for outcome. These scores can then be converted into an assessment of overall study quality as ‘good,’ which requires 3–4 stars in selection, 1–2 stars in comparability and 2–3 stars in outcome; ‘fair,’ which requires 2 stars in selection, 1–2 stars in comparability and 2–3 stars in outcome and ‘poor’, which includes studies with 0–1 stars in selection, or 0 stars in comparability or 0–1 stars in outcome.

If there are any randomised studies eligible for inclusion, these will be assessed for bias using the Cochrane Risk of Bias (RoB) 2 tool.20 The Cochrane RoB 2 tool assesses randomised studies across five domains, which are the randomisation process, deviations from the intended effect, missing outcome data, measurement of outcome and reported results. The risk of bias in each domain is assessed as low, high or some concerns, according to the RoB 2 guidelines. The overall study risk of bias can then be judged as ‘low risk,’ if the study is assessed as low risk in all domains, ‘some concerns,’ if the study is assessed as some concerns in at least one domain, but not high risk in any domain, or ‘high risk,’ if the study is assessed as high risk in at least one domain, or some concerns in multiple domains.

Data analysis

We intend to calculate an effect size (ES) for each study, which is the proportion of patients who achieve a favourable outcome who have localising 18F-FDG-PET hypometabolism compared with those without localising 18F-FDG-PET hypometabolism, as defined above in the intervention. We will then use Der Simonian and Laird random effects meta-analysis to calculate the summary estimate of ES, pooled favourable outcome rate, and 95% CIs. Pooled ES will be presented as the percentage of patients achieving seizure freedom. Statistical heterogeneity of included studies will be measured with I2.

In our secondary analysis, we will stratify studies based on age group (adult, paediatric or both), resected lobe (frontal, insular, temporal, parietal or occipital) and histopathology (hippocampal sclerosis, focal cortical dysplasia types 1 and 2, tumour, vascular malformation or other), and perform random effects meta-analyses on stratified groups. We also intend to perform subgroup analyses, expressed as a risk ratio, of the following variables on favourable outcome: extent of 18F-FDG-PET hypometabolism (focal vs regional or diffuse), location of hypometabolism (temporal vs extratemporal) and proportion of hypometabolism zone resected (<50% vs ≥50%).

If sufficient data are available, from a minimum of five publications, we will perform stratified meta-analysis, using the same methodology used for the primary meta-analysis described above, and meta-regression to investigate and characterise sources of heterogeneity based on the following variables: age group (adult vs paediatric vs mixed), seizure freedom classification (Engel vs ILAE vs other), location (temporal vs extra-temporal), and duration of post-operative follow-up reported (12–23 months, 24–59 months or ≥60 months). The risk of publication bias will be assessed using a funnel plot and Egger’s test.

Patient and public involvement

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Ethics and dissemination

The proposed systematic review does not require ethics approval, as the data to be collected cannot be connected to individual patients. We intend to publish the results of the systematic review in a high quality, peer-reviewed journal.

Discussion

Our proposed systematic review and meta-analysis will evaluate and review published data on the role of localisation with 18F-FDG-PET in predicting favourable outcomes in patients undergoing epilepsy surgery. If localisation with 18F-FDG-PET does predict favourable outcomes in epilepsy surgery, then this systematic review will provide evidence for the inclusion of 18F-FDG-PET localisation in future multimodal outcome prediction algorithms for epilepsy surgery. We anticipate that our systematic review will also help to guide future research into the role of 18F-FDG-PET in epilepsy surgery by further characterising knowledge gaps relating to this topic.

To our knowledge, this will be the first systematic review and meta-analysis addressing the role of localisation with 18F-FDG-PET for patients with all types of drug resistant focal epilepsy who have undergone resective epilepsy surgery. The proposed systematic review and meta-analysis will update and build on previous meta-analyses, which focused on temporal or frontal epilepsies.10 14 15 Wang et al published their meta-analysis in 2016, however, the most recent publication that was included in their meta-analysis of localisation with 18F-FDG-PET to predict outcomes in epilepsy surgery was from 2012.15 Therefore, we believe that undertaking this systematic review and meta-analysis is important and justified, as we expect there will be a considerable number of eligible studies published over the last 10 years.

There are several potential limitations to our proposed study. First, we have adopted a broad definition of 18F-FDG-PET localisation in this study, which considers both the final decision regarding surgical site and concordance with one or more other presurgical diagnostic investigations. While this reflects routine decision-making regarding 18F-FDG-PET, this may be a source of heterogeneity. Second, we have elected to include publications with both randomised and non-randomised study designs, which may lead to the inclusion of low-quality studies. Furthermore, we expect that the majority, if not all, of the included studies will have a non-randomised design, due to the nature of the study question, which are inherently at higher risk of bias than randomised studies. We hope to mitigate this potential risk with a robust assessment of bias as described in our methodology. A third potential limitation is that we are only choosing to include English-language publications, and this may mean that relevant data is missed. Finally, while we hope to have a large enough sample size for our primary meta-analysis, it is possible that our proposed subgroup analyses will be limited by small sample sizes.